Machine Learning

Could the best LLM be able to generate a symbolic AI that is superior to itself, or is there something superior about matrices vs graphs?

Deep neural network AIs have beaten symbolic AIs across the board on many tasks, but is there a chance that symbolic AIs written by DNNs(...

Improving AI models’ ability to explain their predictions

New technique makes AI models leaner and faster while they’re still learning

[2604.03280] Multi-Agent Training-free Urban Food Delivery System using Resilient UMST Network

Abstract page for arXiv paper 2604.03280: Multi-Agent Training-free Urban Food Delivery System using Resilient UMST Network

[2604.03275] IPSL-AID: Generative Diffusion Models for Climate Downscaling from Global to Regional Scales

Abstract page for arXiv paper 2604.03275: IPSL-AID: Generative Diffusion Models for Climate Downscaling from Global to Regional Scales

[2604.03253] Self-Execution Simulation Improves Coding Models

Abstract page for arXiv paper 2604.03253: Self-Execution Simulation Improves Coding Models

[2604.04916] Empowering Power Outage Prediction with Spatially Aware Hybrid Graph Neural Networks and Contrastive Learning

Abstract page for arXiv paper 2604.04916: Empowering Power Outage Prediction with Spatially Aware Hybrid Graph Neural Networks and Contra...

[2604.04908] HI-MoE: Hierarchical Instance-Conditioned Mixture-of-Experts for Object Detection

Abstract page for arXiv paper 2604.04908: HI-MoE: Hierarchical Instance-Conditioned Mixture-of-Experts for Object Detection

[2604.04902] Are Latent Reasoning Models Easily Interpretable?

Abstract page for arXiv paper 2604.04902: Are Latent Reasoning Models Easily Interpretable?

[2604.04892] Data Attribution in Adaptive Learning

Abstract page for arXiv paper 2604.04892: Data Attribution in Adaptive Learning

[2604.04869] Optimizing LLM Prompt Engineering with DSPy Based Declarative Learning

Abstract page for arXiv paper 2604.04869: Optimizing LLM Prompt Engineering with DSPy Based Declarative Learning

[2604.04868] Noise Immunity in In-Context Tabular Learning: An Empirical Robustness Analysis of TabPFN's Attention Mechanisms

Abstract page for arXiv paper 2604.04868: Noise Immunity in In-Context Tabular Learning: An Empirical Robustness Analysis of TabPFN's Att...

[2604.04858] FairLogue: A Toolkit for Intersectional Fairness Analysis in Clinical Machine Learning Models

Abstract page for arXiv paper 2604.04858: FairLogue: A Toolkit for Intersectional Fairness Analysis in Clinical Machine Learning Models

[2604.04855] The Role of Generator Access in Autoregressive Post-Training

Abstract page for arXiv paper 2604.04855: The Role of Generator Access in Autoregressive Post-Training

[2604.04808] Selecting Decision-Relevant Concepts in Reinforcement Learning

Abstract page for arXiv paper 2604.04808: Selecting Decision-Relevant Concepts in Reinforcement Learning

[2604.04800] Forgetting to Witness: Efficient Federated Unlearning and Its Visible Evaluation

Abstract page for arXiv paper 2604.04800: Forgetting to Witness: Efficient Federated Unlearning and Its Visible Evaluation

[2604.04767] Cog-DRIFT: Exploration on Adaptively Reformulated Instances Enables Learning from Hard Reasoning Problems

Abstract page for arXiv paper 2604.04767: Cog-DRIFT: Exploration on Adaptively Reformulated Instances Enables Learning from Hard Reasonin...

[2604.04756] Darkness Visible: Reading the Exception Handler of a Language Model

Abstract page for arXiv paper 2604.04756: Darkness Visible: Reading the Exception Handler of a Language Model

[2604.04717] The Infinite-Dimensional Nature of Spectroscopy and Why Models Succeed, Fail, and Mislead

Abstract page for arXiv paper 2604.04717: The Infinite-Dimensional Nature of Spectroscopy and Why Models Succeed, Fail, and Mislead

[2604.04736] Sampling Parallelism for Fast and Efficient Bayesian Learning

Abstract page for arXiv paper 2604.04736: Sampling Parallelism for Fast and Efficient Bayesian Learning

[2604.04701] MUXQ: Mixed-to-Uniform Precision MatriX Quantization via Low-Rank Outlier Decomposition

Abstract page for arXiv paper 2604.04701: MUXQ: Mixed-to-Uniform Precision MatriX Quantization via Low-Rank Outlier Decomposition

[2604.04698] Explainable Machine Learning for Sepsis Outcome Prediction Using a Novel Romanian Electronic Health Record Dataset

Abstract page for arXiv paper 2604.04698: Explainable Machine Learning for Sepsis Outcome Prediction Using a Novel Romanian Electronic He...

[2604.04681] Batch Loss Score for Dynamic Data Pruning

Abstract page for arXiv paper 2604.04681: Batch Loss Score for Dynamic Data Pruning